What Drives Away Twice as Many Customers as Fraud? A Problem Many Merchants Don't Know They Have.
Just like consumers and business buyers, fraudsters have ramped up their ecommerce activity this year. As retailers struggle to make up for revenue lost to closures and slowdowns in physical stores, rooting out online fraud should be a top priority. Most merchants are aware of the impact of fraud, which is why many are happy to share the fact that they don’t have a fraud problem at all.
But when the number of completed fraud attempts in a store is zero, month after month, it may mean that the automated fraud protection program is so strict that it’s creating another problem: false declines. Depending on how your fraud prevention program classifies rejected orders, you may not have any idea how many good orders are getting rejected by mistake.
If you do have a false decline problem, it may be costing you much more revenue than fraud would. It could be ruining your reputation with customers, too. A consumer survey commissioned by ClearSale in March found that online shoppers in five countries are much more likely to stop shopping with a merchant after a false decline than they are after fraud.
Here’s what every merchant needs to know about the problems that false declines can cause, how to figure out if false declines are a problem for your business, and how to stop them from happening.
Why do false declines happen and why are they hard to uncover?
False declines happen when anti-fraud software mistakes good orders for fraud. Why would an anti-fraud program make that mistake, especially if it’s powered by AI and machine learning?
One reason is that many fraud programs are set to reject orders with any discrepancy that might indicate fraud—such as shipping to a new address or an order from an old customer on a new device. But people move around, ship items to new places and get new phones, so super-strict guidelines can create a lot of false declines. For example, we’ve seen fraud programs that reject every single order with an AVS (address verification system) mismatch even though 90% of the orders were good.
False declines usually don’t show up in merchants’ data because the fraud program labels all rejected orders as fraud attempts, with no follow-up to see if any of those orders were actually good. Because most merchants never see them, false positives give merchants false confidence about the quality of their fraud protection, while their customer base is silently eroding.
False declines drive away twice as many customers as fraud does
In March, ClearSale commissioned a study from Sapio Research on consumer attitudes toward ecommerce fraud in the U.S., U.K., Australia, Canada and Mexico. Sapio collected data from about 1,000 frequent online shoppers in each country.
One notable result was that, on average, consumers are twice as likely to drop a merchant who rejects their order as they are to avoid doing business with a merchant after a fraud experience in their online store.
The strength of this response varies by country: More than half of Mexican shoppers will ditch a brand after a false decline, but only 16% will stop shopping with a brand after fraud. Americans are the most relaxed about false declines, but fully 1/3 of them will abandon a brand after one. In the U.K., Canada and Australia, 38% of shoppers say goodbye after a rejected order.
False declines drive drown revenue, drive up costs, damage brands
This data shows one of the key reasons that false declines are such a big cost to merchants. Not only do stores lose the revenue from the rejected orders, but they also lose the marketing budget they spent bringing those shoppers to their site. If those customers never return, the merchant sees a decrease in customer lifetime value and an increase in the average cost to acquire each customer.
But there are even more costs to false declines. Note the number of consumers who told Sapio they would post something on social media after a false decline. Anywhere from 22% (in Australia) to 39% (in Mexico) of rejected customers will complain. Those posts can damage a merchant’s reputation and make it even harder and more expensive to acquire new customers.
False declines can even make a store’s fraud prevention program less accurate. How? False declines look real to your fraud program’s machine learning, so unless you teach it otherwise it’s going to keep rejecting similar good orders—and the problem will continue to cost you money.
How can you find your false decline rate?
Put on your detective hat, get a team together, and dig into a batch of rejected orders. First, you’ll need to find your average decline rate – the percentage of orders in a given month that are rejected because of suspected fraud.
Then you’ll need to do one of two things. You can have a group of fraud analysts manually review random batches of your rejected orders to see which ones were false declines. You can then use that data to calculate the average percentage of your rejected orders that were really good orders.
If auditing your rejected orders isn’t an option, you can estimate, using industry figures for false declines. These range from 30% to 65% of all rejected orders. In my experience, the higher end of the range is more accurate for many merchants.
Once you know, or have an estimate, of your false decline rate, you can figure out how much revenue you’re losing to false declines. You’ll need to know how much money you lose on orders that are rejected at the payment flow stage and how much money you lose on orders that are rejected by your in-house or third-party fraud prevention solution. Add those two figures together and multiply by your false decline rate.
For example, if you lose an average of $500 each month on orders rejected at the payment flow stage and another $1000 on orders kicked out by your fraud prevention program, you may think you’re blocking $1,500 in fraud losses each month. But if your false decline rate averages 50%, you’re only blocking $750 a month in fraud losses, while losing another $750 in good orders.
And you’re driving away many of those customers forever. If you want to see what that’s costing you in customer lifetime value, you can divide your average revenue lost to false declines by your store’s average order value. Then multiply the result by your average customer lifetime value.
For example, if you’re losing $100,000 a month to false declines, your average monthly order value is $10,000, and your average customer lifetime value is $3,000, you’re also losing an average of $30,0000 per month in customer lifetime value to false declines.
How can you stop false declines, earn more revenue and keep your good customers?
Realize that the rules that applied a year ago for transaction screening can contribute to false declines now. You may need to review and adjust rules about new customers, new online shoppers, new shipping addresses, new devices and new buying behaviors to account for pandemic-related changes in daily life and shopper budgets.
Turn off automatic rejections and manually review flagged orders in-house or through a third party. You can also arrange for a hybrid of in-house and on-call third-party manual review so you can scale up manual review capacity during seasonal sales peaks without draining your budget. Manual review is the best way to prevent good orders from being rejected.
Finally, feed your manual review results back into your fraud machine learning so it gets smarter about separating good orders from fraud attempts. Over time, this should result in fewer good orders getting flagged in the first place, which will result in fewer manual reviews.
The first step to fixing false declines is finding out if you have a problem. Doing the audits and implementing a review system takes time, but it’s worth it to keep your customers, make more sales and maximize your online revenue.